{"id":"W4312516176","doi":"10.1162/tacl_a_00511","title":"Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond","year":2022,"lang":"en","type":"article","venue":"Transactions of the Association for Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":194,"is_retracted":false,"has_abstract":true,"ca_institutions":"Columbia College","funders":"","keywords":"Causal inference; Computer science; Interpretability; Inference; Artificial intelligence; Causality (physics); Natural language processing; Robustness (evolution); Machine learning; Interpretation (philosophy); Data science; Econometrics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.05218536,0.001321264,0.00256068,0.005578053,0.00187628,0.009268486,0.00361209,0.003157155,0.005739526],"category_scores_gemma":[0.2225595,0.001356072,0.002159739,0.005388177,0.01131175,0.01310184,0.005131652,0.007863645,0.0009160257],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003556392,"about_ca_system_score_gemma":0.003908806,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006428292,"about_ca_topic_score_gemma":0.003429423,"domain_scores_codex":[0.9587908,0.0330407,0.001319385,0.003592415,0.002836696,0.000419992],"domain_scores_gemma":[0.6363848,0.3413191,0.007009892,0.009663693,0.004891913,0.000730554],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.000126952,0.0000830975,0.007146084,0.0009523456,0.000466533,0.0002491954,0.001044122,0.03859647,0.0003530164,0.8501629,0.006194159,0.09462497],"study_design_scores_gemma":[0.00001336968,0.00000948745,0.0005197817,0.000228905,0.00003448012,0.00004119903,0.00006605686,0.0632639,0.0001482387,0.9326429,0.003011555,0.00002015631],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.006548898,0.008914301,0.9654197,0.0140366,0.0002423743,0.00009624548,0.0004610413,0.0004408295,0.003840042],"genre_scores_gemma":[0.6168602,0.01481231,0.3547273,0.004995914,0.003607524,0.0006780988,0.001372011,0.0004345402,0.002512107],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.05218536,"threshold_uncertainty_score":0.2759859,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.008751666553274558,"score_gpt":0.2654126314121237,"score_spread":0.2566609648588491,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}